Upadacitinib plus topical corticosteroids in atopic dermatitis: Week 52 AD Up study results

Jonathan I. Silverberg, MD, Marjolein de Bruin-Weller, MD, Thomas Bieber, MD, Weily Soong, MD, Kenji Kabashima, MD, Antonio Costanzo, MD, David Rosmarin, MD, Charles Lynde, MD, John Liu, MD, AmyGamelli, PhD, Jiewei Zeng, PhD, Barry Ladizinski, MD, Alvina D. Chu,MD, and Kristian Reich, MD Washington, DC; Utrecht, Netherlands; Bonn and Hamburg, Germany; Birmingham, Ala; Kyoto, Japan; Milan, Italy; Boston, Mass; Toronto and Markham, Ontario, Canada; and North Chicago, Ill

52 weeks; no new important safety risks beyond the current label were observed. No deaths were reported; major adverse cardiovascular events and venous thromboembolic events were infrequent (< _0.2/100 patient-years). Conclusions: Results through 52 weeks demonstrate long-term maintenance of efficacy and a favorable safety profile of upadacitinib 1 TCS in patients with moderate-to-severe AD. (J Allergy Clin Immunol 2022;149:977-87.) Key words: Atopic dermatitis, randomized clinical trial, upadacitinib, topical corticosteroids, Janus kinase inhibitors Atopic dermatitis (AD) causes long-term skin-related disability and burden. 1-6 AD onset, most often occurring in childhood before 5 years of age, 7 can occur at any age; symptoms of AD can persist, reemerge, or worsen (flare) throughout a patient's lifetime. 8,9 Long-term persistence of AD is more likely in patients with moderate-to-severe disease. 10 Most patients with AD report symptoms and continue medication use over 2 to 3 decades of life-an important consideration for effective disease management. 11 There is a need for additional AD treatments that are acceptable for long-term use and that provide prolonged clinical response without high levels of treatment discontinuation due to adverse effects. 12,13 Systemic therapies are often used in combination with topical corticosteroids (TCS) to control moderate-to-severe AD symptoms. 14 Adding TCS to dupilumab (anti-IL-4 and -13 receptor alpha monoclonal antibody), tralokinumab (anti-IL-13 monoclonal antibody), or baricitinib (selective Janus kinase [JAK] 1 and JAK2 inhibitor) increases response rates over that observed for monotherapy in patients with moderate-to-severe AD. [15][16][17][18][19][20] Upadacitinib, an oral JAK inhibitor with greater inhibitory potency for JAK1 than JAK2, JAK3, or tyrosine kinase 2, is approved in the United States, the European Union, and other countries to treat moderately or severely active rheumatoid arthritis (RA) 21 and is in development for the treatment of immune-mediated inflammatory conditions, including AD. 22,23 Results from 2 ongoing phase 3, randomized, double-blind, replicate studies (Measure Up 1 [NCT03569293] and Measure Up 2 [NCT03607422]) demonstrate superiority of upadacitinib 15 mg and upadacitinib 30 mg versus placebo (PBO). As monotherapy, both upadacitinib doses are well tolerated; markedly improve skin signs, itch, skin pain, and health-related quality of life; and consistently exhibit higher thresholds of skin improvement (ie, > _90%/ > _100% improvement in Eczema Area and Severity Index [EASI-90/EASI-100]) through 16 weeks in adolescents and adults with moderate-to-severe AD. 24 Primary results from AD Up, a pivotal, phase 3, randomized, double-blind study, provide evidence that upadacitinib 1 TCS is well tolerated and superior to PBO 1 TCS in adolescents and adults with moderate-to-severe AD; significantly greater proportions of patients receiving either upadacitinib dose plus TCS experience > _75% improvement in EASI (EASI-75), EASI-90, and EASI-100 and validated Investigator's Global Assessment of AD (vIGA-AD) 25 of clear or almost clear with > _2 grades of improvement (vIGA-AD 0/1) compared with PBO 1 TCS. 26 No new important safety signals have been observed in the phase 3 AD program beyond those reported in the RA clinical program, 21 demonstrating an overall favorable benefit-risk profile of upadacitinib in AD. Exploration of the long-term benefit-risk profile of adding standard TCS to upadacitinib continues in the ongoing blinded extension (BE) period of AD Up; here we report results through 52 weeks.

Study design and patients
Study design and patient population details have been previously reported. 26 Briefly, AD Up (NCT03568318) was a pivotal, phase 3, randomized, double-blind, PBO-controlled, multicenter study conducted in 171 clinical centers globally, consisting of a main study (reported here) and an adolescent substudy (ongoing; to be reported elsewhere). The study had a 35-day screening period and 16-week double-blind period (reported previously), 26 followed by a BE period for up to 260 weeks of treatment with a 30-day follow-up (ongoing; results up to week 52 reported here).
Independent ethics committees/institutional review boards at each study site approved the study protocol, informed consent forms, and recruitment materials before patient enrollment. The study was conducted in agreement with the International Conference for Harmonization guidelines, applicable regulations, and the Declaration of Helsinki. Adult patients and parents/legal guardians of adolescent patients provided written informed consent before any screening or study-related procedures. The adolescent substudy was added after protocol initiation to allow enrollment of additional adolescents to fulfill a regulatory commitment.
With the advent of the coronavirus disease 2019 (COVID-19) pandemic, operational accommodations for clinical trial continuity were incorporated for temporary site disruptions and secure-in-place measures (see the Methods in this article's Online Repository at www.jacionline.org).

Randomization and masking
Eligible patients were randomized 1:1:1 using an interactive response technology system according to a schedule generated by statisticians at AbbVie to receive upadacitinib 15 mg, upadacitinib 30 mg, or PBO once daily, all in combination with TCS. At week 16, PBO 1 TCS-treated patients were rerandomized 1:1 to receive upadacitinib 15 mg or 30 mg plus TCS; patients initially randomized to upadacitinib 15 mg 1 TCS or upadacitinib 30 mg 1 TCS continued treatment as originally assigned. Randomization in the main study was stratified by baseline vIGA-AD score (moderate, 3; severe, 4), geographic region (United States/Puerto Rico/Canada, Japan, China, Other), and age group (adolescent, adult). Rerandomization was stratified by EASI-50 (> _50% improvement from baseline in EASI) response at week 16, age group, and region for the BE up to week 52. Study investigators, study site personnel, and patients remained unaware of their treatment regimen throughout the study; the upadacitinib 15 mg, upadacitinib 30 mg, and PBO tablets were identical in appearance.

Procedures
Patients took a single oral tablet of upadacitinib 15 mg, upadacitinib 30 mg, or PBO (AbbVie, North Chicago, Ill) once daily. Twice-daily use of an additive-free, bland emollient was required for > _7 days before baseline and during the study until week 52. Protocol-mandated TCS was applied through week 52 according to a step-down regimen described previously. 26 Briefly, medium-potency TCS (low-potency TCS or topical calcineurin inhibitor for sensitive skin areas) was applied once daily to areas with active lesions for < _3 consecutive weeks or until lesions were clear or almost clear; then low-potency TCS was applied once daily for 7 days and stopped if lesions were no longer active (for sensitive skin areas, low-potency TCS or topical calcineurin inhibitor was tapered and stopped). If lesions returned or persisted, this step-down approach was repeated until lesion resolution or evidence of local or systemic TCS toxicity (eg, striae, skin atrophy, bruising). Although the TCS potency was mandated, the selection of TCS within each potency category was per investigator choice; no specific TCS was mandated. As-needed therapy was not permitted, but starting at week 4, rescue therapy was permitted based on lack of EASI response. Starting at week 52, TCS use was per investigator discretion (see the Methods in this article's Online Repository).

Outcomes
The co-primary efficacy end points (multiplicity controlled) were the proportion of patients who experienced EASI-75 and the proportion of patients who experienced vIGA-AD 0/1, both at week 16. The primary and following key secondary efficacy end points were assessed at all visits through week 52: the proportions of patients who experienced WP-NRS improvement > _4, EASI-90 and EASI-100, and the percentage changes from baseline in EASI and WP-NRS. Among those with response, defined as patients experiencing vIGA-AD 0/1 and EASI-75 at week 16, the proportion of patients experiencing a loss of response after week 16 until week 52 was assessed by visit and overall; loss of response was defined as a loss of > _50% of the EASI response at week 16 and vIGA-AD score > _2 after week 16.
Safety data were presented as of the data cutoff for the week 52 analysis. The following safety parameters were assessed: treatment-emergent adverse events (TEAEs); serious adverse events; deaths; adverse events (AEs) leading to discontinuation of study treatment; prespecified AEs of special interest, which were based on the known upadacitinib safety profile 21 and previous safety observations for upadacitinib 23 and other JAK inhibitors 28 in patients with AD; laboratory assessments; and vital signs.

Statistical analysis
A sample size of 810 patients (270 per treatment group) was estimated to provide >90% power to detect treatment differences for the 2 primary end points of 38% and 20%, respectively, and to allow for adequate characterization of safety. 26 After all patients in the main study completed the therapy regimen at week 52, efficacy analyses were conducted in the intent-to-treat population for the main study, defined as all patients who were randomized into the main study. Safety analyses were conducted in the safety population for the main study up to the cutoff date of the week 52 analysis, defined as all randomized patients who received > _1 dose of study drug. PBO 1 TCS-treated patients rerandomized at week 16 were combined into their respective upadacitinib 15 mg 1 TCS or upadacitinib 30 mg 1 TCS groups for the safety analyses.
For categorical end points, the primary approach for handling missing data was nonresponse imputation incorporating multiple imputation (MI) for missing values due to COVID-19 (NRI-C). 26 For response rates, 95% CIs were based on a Student t distribution from the SAS PROC MIANALYZE procedure (SAS Institute, Cary, NC). NRI-C was the primary approach for categorical end points in the double-blind period and for co-primary end points and WP-NRS in the BE period up to week 52. MI was also performed as a confirmatory sensitivity analysis for EASI-75/-90/-100, vIGA-AD 0/1, and WP-NRS in the BE period up to week 52 using Markov chain Monte Carlo and PROC MI, and included the following: treatment group, major stratum (vIGA-AD categories [for end points other than vIGA-AD 0/1], age group, and region), gender, and measurements at baseline and each visit up to the end of the analysis period. All assessments after the start of rescue medication were not included in the analyses; patients were counted as having nonresponsive disease after receiving rescue medication, and data were not imputed by MI (see the Methods in this article's Online Repository).
For continuous end points, change and percentage change from baseline in each treatment group were analyzed using the mixed-effects model for repeated measures. This included categorical fixed effects of treatment, visit, and treatment-by-visit interaction; the continuous fixed covariates of baseline measurement; and stratification factors of vIGA-AD categories at randomization (moderate vs severe) and age group for the double-blind period, and EASI-50 response at week 16 and age group for the BE period up to week 52. Observed case analysis, which did not impute values for missing evaluations, was used to assess loss of response among those with response at week 16 and was performed for all variables in addition to NRI-C and MI. Further, post hoc analyses for percentage change of EASI and WP-NRS were conducted for the upadacitinib treatment groups from baseline through week 52 with treatment (upadacitinib 15 mg and upadacitinib 30 mg), treatment-by-visit interaction, vIGA-AD categories at baseline, age (adolescent vs adult), and baseline value in the model. Using descriptive statistics only, safety data were reported as the number of AEs divided by the total exposure in 100 patient-years (PY), and potentially clinically significant changes from baseline in laboratory assessments were reported as the number and proportion of patients.

RESULTS
Between August 9, 2018, and December 20, 2019, a total of 901 patients were randomized to the double-blind period (300 to upadacitinib 15 mg 1 TCS, 297 to upadacitinib 30 mg 1 TCS, and 304 to PBO 1 TCS). As previously reported, demographics and baseline disease characteristics were well balanced between groups. 26 At week 16, a total of 283 PBO 1 TCS-treated patients were rerandomized-144 to upadacitinib 15 mg 1 TCS (143 were treated) and 139 to upadacitinib 30 mg 1 TCS (139 were treated)-while 289 upadacitinib 15 mg 1 TCS-treated and 287 upadacitinib 30 mg 1 TCS-treated patients continued to receive their initial treatment in the BE period (Fig 1). During the BE period, no patients discontinued treatment for primary reasons related to COVID-19. The number of patients for whom missing EASI-75 and/or vIGA-AD 0/1 data were imputed using MI due to COVID-19 are listed in Table E1 in this article's Online Repository at www.jacionline.org. Rescue medication was initiated in the BE period by 15.0%, 8.1%, 4.2%, and 5.0% of patients receiving upadacitinib 15 mg 1 TCS, upadacitinib 30 mg 1 TCS, PBO 1 TCS/upadacitinib 15 mg 1 TCS, and PBO 1 TCS/upadacitinib 30 mg 1 TCS, respectively.
Overall, similar results were obtained with NRI-C, MI, and observed case analyses (see Figs E4 to E11).
As of the 52-week data analysis cutoff, both upadacitinib 15 mg 1 TCS and upadacitinib 30 mg 1 TCS were well tolerated ( Table I). The exposure-adjusted event rates (E/100 PY) of any TEAE, serious AEs, and AEs leading to discontinuation of study drug were similar between upadacitinib 15 mg 1 TCS and upadacitinib 30 mg 1 TCS groups (Table I). No deaths were reported through 52 weeks of the BE period (Table I).
The most frequently reported TEAEs (> _10% in either treatment group) were acne, nasopharyngitis, blood creatine phosphokinase (CPK) increase, dermatitis atopic, and upper respiratory tract  infection (Table I). No acne events were serious, and 1 mild acne event in the upadacitinib 30 mg 1 TCS group led to study drug discontinuation on study day 22. Rates of serious infections were similar between treatment groups (2.7 and 2.3 E/100 PY with upadacitinib 15 mg 1 TCS and upadacitinib 30 mg 1 TCS, respectively) ( Table I). All opportunistic infections reported, excluding tuberculosis and herpes zoster (HZ), were cases of eczema herpeticum (Kaposi varicelliform eruption); the exposure-adjusted event rates were 2.7 and 2.3 E/100 PY in the upadacitinib 15 mg 1 TCS and upadacitinib 30 mg 1 TCS treatment groups, respectively. All eczema herpeticum events were nonserious, and 2 events reported in the upadacitinib 30 mg 1 TCS group lead to treatment discontinuation. The event rate of HZ was higher in the upadacitinib 30 mg 1 TCS group compared with upadacitinib 15 mg 1 TCS (6.2 vs 4.1 E/100 PY). Most HZ events involved a single dermatome and did not lead to treatment discontinuation. There were no HZ events involving central nervous system, lung, or liver. One squamous cell carcinoma of the oral cavity was reported in the upadacitinib 15 mg 1 TCS group; 3 malignancies were reported in the upadacitinib 30 mg 1 TCS group, with 1 case each of nonmelanoma skin cancer (a keratoacanthoma), colon adenocarcinoma, and melanoma-insitu of the digit. Colon adenocarcinoma and keratoacanthoma  were diagnosed less than 2 months from the first dose of upadacitinib. Adjudicated major adverse cardiovascular events were reported for a 60-year-old patient treated with upadacitinib 15 mg 1 TCS (nonfatal subarachnoid hemorrhage) and a 69year-old patient treated with upadacitinib 30 mg 1 TCS (nonfatal stroke); both were classified as serious events and were deemed unrelated to study drug with cardiovascular risk factors, and the patients were withdrawn from study treatment. One adjudicated venous thromboembolic event (venous thromboembolic event [grade 2 pulmonary embolism]) was reported for a 66-year-old white male patient with a history of obesity and hypercholesterolemia in the upadacitinib 15 mg 1 TCS group. This event was an incidental finding based on results from a routine chest X-ray specified by the protocol for the week 52 visit, and was deemed not serious and possibly related to study drug by the study physician. The patient was withdrawn from the study as a result of this event. There were no reports of active tuberculosis, adjudicated gastrointestinal perforation, or lymphoma.
Most AEs of hepatic disorders were transient, asymptomatic transaminase elevations. The event rates of alanine aminotransferase and aspartate aminotransferase increased were 3.3 and 1.8 E/100 PY with upadacitinib 15 mg 1 TCS and 1.5 and 1.3 E/ 100 PY with upadacitinib 30 mg 1 TCS, respectively. AEs of anemia, neutropenia, and CPK elevations were reported more frequently with upadacitinib 30 mg 1 TCS versus upadacitinib 15 mg 1 TCS. These laboratory-related AEs were generally nonserious and did not lead to treatment discontinuation. Most CPK elevations occurred after exercise or other vigorous physical activity (76.6% in upadacitinib 15 mg 1 TCS and 64.0% in upadacitinib 30 mg 1 TCS groups); most (> _84%) did not have any associated symptoms. Overall, potentially clinically important laboratory test results were infrequent (Table II). The incidence of grade 3 or higher elevations in CPK showed a dose-related increase with upadacitinib treatment.

DISCUSSION
This report provides the first evidence of long-term efficacy and safety of upadacitinib 1 TCS through 52 weeks of treatment. Although there were no statistical comparisons between doses, a clear dose response was observed from week 2 through week 52: patients treated with upadacitinib 30 mg 1 TCS consistently experienced numerically better results compared with upadacitinib 15 mg 1 TCS. This trend was also observed in PBO 1 TCStreated patients who were rerandomized to receive upadacitinib 15 mg 1 TCS or upadacitinib 30 mg 1 TCS from week 16 to week 52. Upadacitinib 1 TCS provides meaningful clinical responses (ie, vIGA-AD 0/1 and EASI-75) as well as extensive responses (ie, EASI-90 and EASI-100) in adolescents and adults with moderate-to-severe AD. Notably, no new important safety risks were observed through 52 weeks of treatment beyond those described in the current label for RA. 21 The only laboratoryrelated AE anomaly was CPK elevations, which were generally nonserious, were related to vigorous physical activity, and did not lead to treatment discontinuation. Safety results were similar between upadacitinib doses, with clinically irrelevant differences between the upadacitinib 15 mg 1 TCS and upadacitinib 30 mg 1 TCS groups. These results reinforce the primary (week 16) efficacy and safety results 26 and demonstrate efficacy and favorable safety of upadacitinib 1 TCS is maintained long term through 52 weeks of treatment.
There is an overall lack of long-term efficacy and safety data for systemic-plus-TCS treatments in patients with AD. CHRONOS was the first randomized, double-blind, PBO-controlled, phase 3 study of long-term (1-year) systemic dupilumab treatment in combination with TCS in patients with moderate-to-severe AD and inadequate response to TCS. Primary (week 16) improvements with dupilumab 1 TCS were sustained over 52 weeks of treatment. At week 52, significantly more patients who received dupilumab 1 TCS experienced vIGA-AD 0/1 (36% [n 5 32] with dupilumab 300 mg every 2 weeks 1 TCS and 40% [n 5 108] with dupilumab 300 mg once weekly 1 TCS vs 13% [n 5 33] with PBO 1 TCS) and EASI-75 (65% [n 5 58] with dupilumab 300 mg every 2 weeks 1 TCS and 64% [n 5 173] with dupilumab 300 mg every week 1 TCS vs 22% [n 5 57] with PBO 1 TCS) (P < .001 for all). Several studies of long-term efficacy and safety of JAK inhibitors in combination with TCS to treat moderate-to-severe AD are ongoing. 19,20,23,29 Though preliminary short-term results are promising, limited data are available for long-term efficacy and safety of JAK inhibitors in combination with TCS to treat moderate-to-severe AD. The 52-week results  J ALLERGY CLIN IMMUNOL VOLUME 149, NUMBER 3 from AD Up reported here provide the first evidence of long-term efficacy and acceptable safety profile of upadacitinib in AD. Compared with the RA program, 21 no new important safety risks were observed; however, the rate of acne was higher in the AD study. These acne events were nonserious and rarely led to treatment discontinuation.
Results obtained with each individual imputation method provide different and relevant information that, when taken together, may inform the clinical use of the drug. The primary NRI-C analysis is a stringent analysis that cumulatively applies nonresponse forward into subsequent time points; the accumulation of nonresponse over time could contribute to a downward response rate trend. PBO 1 TCS-treated patients who switched to upadacitinib 15 mg 1 TCS or upadacitinib 30 mg 1 TCS had high response rates (based on those who received > _1 dose of study drug in the BE period) from weeks 20 to 52. This population may have been enriched for patients who made it through the PBOcontrolled period without rescue medication or EASI score worsening of > _25% at any 2 consecutive visits.
EASI and WP-NRS percentage change appeared to rebound after week 16, which may be related to the analysis model. In the prespecified analysis for the percentage change of EASI and WP-NRS, separate models were used for the double-blind period and the BE period up to week 52 as a result of changes in stratification factors and number of treatment groups. Results from a post hoc analysis using a single model for baseline through week 52 did not show the striking rebound from week 16 to week 20 for the percentage change of EASI, while this increase was still observed for upadacitinib 15 mg from week 16 to week 20 for percentage change of WP-NRS. Of note, WP-NRS was administered on electronic handheld devices from screening through week 16; starting at the week 20 visit, the frequency of administration was reduced from daily assessments to assessments only at scheduled site visits using a tablet at the site, which may have contributed to the rebound of LSM percentage change in WP-NRS after week 16. Further, if these rebounds in efficacy were meaningful and a true reflection of the drug's efficacy changes between week 16 and week 20, this would also have been captured by other measures based on response analyses. This was not the case, as this did not occur in different thresholds of EASI response (EASI-75/-90/-100), vIGA-AD 0/1, and WP-NRS > _4. Additionally, no patient in any group experienced loss of response at week 20 (Table E2).
Limitations of this analysis include the relatively small sample size and lack of powered statistical comparison between groups from weeks 20 to 52. Given the inherent challenges of studying long-term outcomes in chronic diseases, these results must be examined in context. Also, efficacy results reported here are mostly based on objective outcomes (ie, physician assessment of disease severity) versus subjective outcomes (ie, patient-reported outcomes and quality-of-life assessments).
Future analyses will investigate the long-term impact of upadacitinib 1 TCS on patient-reported outcomes and healthrelated quality of life, rescue medication use, TCS-free days, and disease flare incidence during the BE period. Generalizability of AD Up results compared with other AD studies (eg, CHRO-NOS) is limited by the lack of a PBO 1 TCS treatment group from weeks 16 to 52. The AD Up BE is ongoing; analyses of results for efficacy and safety of upadacitinib 1 TCS through 260 weeks are planned, as well as subgroup analyses (eg, age group). Though upadacitinib 1 TCS closely mimics real-world treatment paradigms for patients with moderate-to-severe AD, upadacitinib 1 TCS experienced similar efficacy for the same end points (eg, skin improvement, itch reduction) with upadacitinib monotherapy in the short term. 24 Therefore, it will be important to compare long-term efficacy and safety of upadacitinib 1 TCS with the long-term upadacitinib monotherapy study efficacy and safety findings (report forthcoming).
In conclusion, results through 52 weeks from the phase 3 AD Up study further support the potential of upadacitinib 1 TCS as a well-tolerated and effective long-term treatment option with a positive benefit-risk profile in adults and adolescents with moderate-to-severe AD.

Data sharing
AbbVie is committed to responsible data sharing regarding the clinical trials we sponsor. This includes access to anonymized, individual, and trial-level data (analysis data sets), as well as other information (eg, protocols and clinical study reports), as long as the trials are not part of an ongoing or planned regulatory submission. This includes requests for clinical trial data for unlicensed products and indications. These clinical trial data can be requested by any qualified researchers who engage in rigorous, independent scientific research, and will be provided following review and approval of a research proposal and statistical analysis plan and execution of a data sharing agreement. Data requests can be submitted at any time, and the data will be accessible for 12 months, with possible extensions considered. For more information on the process, or to submit a request, visit the following link: https://www.abbvie.com/our-science/clinical-tri als/clinical-trials-data-and-information-sharing/data-and-inform ation-sharing-with-qualified-researchers.html.
We thank all study investigators for their contributions and the patients who participated in this study. AbbVie funded the research for this study and provided writing support for this article by funding the medical writing assistance provided by Jennifer C. Jaworski, MS, and Caroline W. Cazares, PhD, of JB Ashtin.

Protocol-mandated TCS
The sponsor did not provide TCS, and choice of TCS aligned with potency was at the investigator's discretion. The protocol recommended triamcinolone acetonide 0.1% cream or fluocinolone acetonide 0.025% ointment as mediumpotency TCS and hydrocortisone 1% cream as low-potency TCS.

Rescue therapy
From weeks 4 to 24, if a patient did not experience EASI-50 at any 2 consecutive scheduled study visits, rescue therapy with high-or superhighpotency TCS (unless higher-potency TCS was considered unsafe) or other topical AD medications (eg, topical calcineurin inhibitor or crisaborole) was allowed as needed, escalating to systemic rescue medication only for disease that did not respond adequately after 7 days of topical treatment. From weeks 24 to 52, rescue therapy was allowed as needed if EASI-50 was not recorded at any scheduled visit. Patients who received any rescue therapy were considered to have nonresponsive disease for subsequent visits; however, patients who received rescue therapy with topical AD treatments or oral corticosteroids for < _2 consecutive weeks could continue to receive study medication. Through week 52, rescue therapy was defined as any of the following: high-potency TCS; other topical therapies (not including moisturizers or emollients); biologic, nonbiologic, or other systemic therapies; or phototherapy.

COVID-19 operational accommodations
Measures included remote visits, local laboratory collections, and delivering study drugs to study participants via courier where allowed, in accord with local regulations. Remote assessments of the skin to determine efficacy were not allowed, and in-person visits were required at baseline and week 16.       ITT, Intent to treat for main study. Response was defined as experiencing vIGA-AD 0/1 with > _2 grades of reduction from baseline and EASI-75 at week 16; loss of response was defined as loss of at least 50% of the week 16 EASI response and vIGA-AD score of > _2 or higher.